Hybrid - GenAI Developer || Washington D.C

[Ashu]

HI,

Please suggest if you are good with below role

 

Job Title:             AI Developer

Location:            Washington D.C. (Hybrid, 4 days in a week)

Duration:            Long-term Contract

Pay rate- $50-55/hr c2c

Responsibilities:

• Design, implement, and operate Retrieval-Augmented Generation (RAG) services using Azure AI/Search, including chunking, embeddings, re-ranking, evaluation, and citation display.
• Design and deploy Model Context Protocol (MCP) tools/servers to integrate security scanners, inventory systems, approvals, and Azure DevOps/GitHub services.
• Build agentic AI solutions using AutoGen, CrewAI, and/or Agno, enabling secure tool-calling and multi-agent orchestration for troubleshooting and workflow automation.
• Develop production-grade chatbots (multi-turn, retrieval-grounded) with prompt management, guardrails, audit logging, and telemetry.
• Integrate Azure OpenAI securely behind API Management (APIM), manage secrets with Key Vault, handle events via Event Hub, and instrument with App Insights/Log Analytics.
• Evaluate and (where appropriate) fine-tune open-source models (e.g., PEFT/LoRA), balancing quality, latency, cost, and safety.
• Ship with CI/CD on Azure DevOps, implement unit/integration tests, red-team for prompt-injection/jailbreaks, and document runbooks.

Minimum Qualifications

• 4+ years total software development experience, with 2+ years in applied LLM/GenAI.
• Strong Python skills and hands-on experience with Azure OpenAI and Azure AI/Search (vector search, hybrid search, semantic ranking).
• Practical experience with agent frameworks (AutoGen, CrewAI, Agno) and MCP/tool-use patterns.
• Proven Azure PaaS experience: Azure Functions or Web Apps, APIM, Key Vault, Event Hub; familiarity with Entra ID/RBAC and secure API design.
• Experience implementing observability (App Insights, Log Analytics/KQL) and CI/CD with Azure DevOps.

Nice to Have (including Certifications)

• RAG evaluation frameworks (e.g., Ragas), custom golden sets, KQL proficiency, Cosmos DB familiarity.

Security-first mindset: content safety, prompt-injection defenses, data privacy controls, and threat modeling for AI systems